High-performance open-source vector database Qdrant today announced it’s launching a new managed hybrid cloud model of its database to power enterprise artificial intelligence apps under the Qdrant Cloud as part of its self-service, run-anywhere solution.

Generative AI models, such as the highly popular ChatGPT chatbot from OpenAI, require large swaths of unstructured data such as text, images, audio and video to give them the context for their underlying conversational capabilities. These capabilities include question-and-answer exchanges, search and summarization, text generation and more. Simple keyword searches don’t work with unstructured data, since it’s complex, highly dimensional and uses the contextual relationships between the words, phrases and other elements of information.

Vector databases allow large language models, the AI models that chatbots are built on, to remember context and relationships over time, giving them a memory. This is also necessary to improve the reliability and accuracy of AI models for reducing “hallucinations,” or their tendency to produce errors.

“The technology is pretty new and many companies don’t have the knowledge to approach and scale it,” co-founder and Chief Executive André Zayarni told SiliconANGLE in an interview. Using the hybrid cloud solution companies can keep their data in their private cloud, or another provider, and maintain full control over the data while still getting the benefits of using a vector database.

“Companies are saying, ‘So we have a lot of data, but do we want to push into some third-party solution there in the cloud that we don’t know?’ And that’s what we’re solving with this new feature,” Zayarni said.

Qdrant Hybrid Cloud ensures complete database isolation, which allows customers to safely deploy a vector database in any chosen environment all without giving up on a managed service. Zayarni said this gives customers the best of both worlds without sacrificing the benefits of either, giving customers control over their data and giving them vector search capabilities to power their AI apps.

Because Qdrant is an open-source vector database it can be launched in any environment or cloud provider and it is also supported in any Kubernetes cloud environment and a large number of platforms, including Google Cloud Platform, Amazon Web Services and Microsoft Azure. Zayarni said there are essentially “no limitations” when it comes to where customers can put their data.

He said the important thing is that if a company wants to have a managed version, the company does not need to hand over its data to Qdrant. That will be important for any enterprise customers where proprietary data is being stored in vector databases to create accurate answers from LLMs, but at the same time, the customer does not want that information potentially leaking.

Now that Qdrant has opened up a managed hybrid cloud solution, Zayarni said that the company has plans to look into the potential of offering a serverless vector database solution in the future, which he mentioned is “completely the opposite” of a hybrid solution. As in that case, the service would be offered entirely through Qdrant on an as-used basis.

Image: Pixabay